Upstream Tech has been selected to receive a Small Business Innovation Research (SBIR) Phase 1 award from the Department of Energy (DOE) to support improved hydropower decision-making with the further development of HydroForecast. The original announcement from the DOE’s Water Power Technologies Office is linked here.
Upstream Tech is a US-based public benefit corporation that uses satellite imagery and machine learning to create environmental decision-support technologies. Their HydroForecast service is a neural network-based approach to forecasting streamflow for hydropower operational decision-making. Hydropower operators rely on forecasts for planning and operational decision-making, but the inaccuracy of current forecasts has implications for revenue, risk, and regulatory requirements. Upstream Tech’s SBIR proposal, entitled “Hydropower Decision-Support with Machine Learning and Satellite Driven Forecasts,” leverages machine learning and satellite imagery innovations to improve the accuracy and reliability of hydrologic forecasts.
The Phase 1 project leverages machine learning and satellite imagery innovations to improve the accuracy and reliability of hydrologic forecasts. This work includes three main components:
- Creating a spatially distributed machine learning model to better represent the varied weather and hydrologic processes within a basin;
- Incorporating an ensemble of weather forecasts to more accurately convey the river forecast variability caused by the range of possible future weather patterns; and
- Leveraging the latest research on explainable machine learning to enable users to understand what is driving changes in the forecast and how they make their predictions, via an easy-to-use web dashboard.
